intents
intents copied to clipboard
home-assistant
"Nevermind" should cancel and take priority over all other context of the sentence and fully cancel the request
When "nevermind" is spoken, ideally the fully sentence would be scrapped and considered "cancelled". At the moment this isn't a major problem, because appending "nevermind" to a valid sentence will result in a sentence that can't be fully understood. This results in the typical "Sorry, I couldn't understand that" response.
Not sure if partial understanding of a sentence will be supported in the future, but if so: we'll likely want "nevermind" to always throw the sentence out.
Example scenario:
Note: This could be a non-issue, so feel free to close it if so, just something I noticed. Thanks!
"OK Nabu, play Nevermind by Nirvana" :P
It's an easy fix if it gets enough votes. But I don't support the implementation myself, I haven't seen it implemented in any voice assistant so far.
What if it had priority only if it's the last word(s) uttered? "OK Nabu, play Nevermind by .... uhh forget it..." I think Alexa, Siri and Google Assistant all actually have "nevermind" implemented as well working like above?
As @salleq mentioned above, Alexa focuses on the trialing command: "OK Nabu, play Nevermind by Nirvana. Nevermind." would result in a cancelled request from what I am seeing.
Focusing on the trailing command is a solution that works for me - but, if it's more complicated to implement, then I would also accept cancelling the request if the cancel keyword is spoken at all. It's one word and in most use cases isn't going to be a problem. That word can already be changed by the user. If you're commonly using "nevermind" to not cancel the request, you could change that keyword to be something more applicable to your use case.
I'm fine with whatever the devs/community think is best. If leaving it "as is" is the solution, that works too. Just thought it was worth highlighting. I appreciate the feedback and input!
What if it had priority only if it's the last word(s) uttered? "OK Nabu, play Nevermind by .... uhh forget it..." I think Alexa, Siri and Google Assistant all actually have "nevermind" implemented as well working like above?
I think it would make more sense to just not process "nevermind" for media playing commands, since there's a lot of variability in input.
